es from both ends of the connection,
it's easy to identify firewalls or other interference.
Alan Altmark
z/VM Development
IBM Endicott
--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to lists...@vm
ked by a slow target LPAR. The more transactions you have,
the more MIPS you suck out of the machine. So you have to have the
headroom within all LPARs and the overall CPU capacity to start using
HiperSockets in a big way.
Alan A
OFF LIST. Thanks.
Regards,
Alan
Alan Altmark
Sr. Software Engineer
IBM z/VM Development
--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to lists...@vm.marist.edu with the message: INFO LINUX-390
d my z/OS brethren have been exploring the possibility of going to
RACF for NIS schema data, but at this time the capability you are asking
for is not available.
If that's something you'd like to see, I encourage you to open a
Requirement, whether through the Support Center or your fav
not from
RACF.
Alan Altmark
z/VM Development
IBM Endicott
--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit
http://www.marist.edu/ht
On Monday, 04/26/2010 at 05:11 EDT, Tom Duerbusch
wrote:
> But in z/VM 5.2, support was dropped for IUCV and VCTCA on an IP
network.
Huh? z/VM still continues to support IUCV and CTC (virtual or otherwise)
IP connections. Heck, we still support HYPERchannels and CLAW.
Alan Altmark
z
ne meet with a similiar issue? I don't
> even know what should I check to find the reason of the problem.
You didn't say what kind of error you got. Make sure both side are
reporting chpid 0.
The portname option is not the problem, and you don't want to specify it
65A 665B 665C 665D 665E 665F
Path 53 online to devices 67E1
Ready; T=0.01/0.01 23:20:08
q chpid 53 type
Path 53 is an Open Systems Adapter Direct Express (OSD) channel
Ready; T=0.01/0.01 23:20:16
Alan Altmark
z/VM Development
IBM Endicott
---
get the
chpid number on the working LPAR. Then you can use 'lschp' to ensure that
the chpid is shared. On the failing LPAR, use lsqeth again to ensure you
have the same chpid.
Alan Altmark
z/VM Development
IBM Endicott
--
On Sunday, 05/09/2010 at 11:46 EST, Błażej Ślusarek
wrote:
> Seems strange to me that chpid from lsqeth is not present in lschp (or
> I don't understand the output).
Sounds like a bug to me.
Alan Altmark
z/VM Development
I
On Monday, 11/01/2004 at 08:42 MST, "Seader, Cameron"
<[EMAIL PROTECTED]> wrote:
> Why is that? i was able to do 64k on SLES 8
Actually, Cameron, you were not. HiperSockets are physically incapable of
handling an MTU larger than 56K. You may have *specified* 64K, but you
didn
o drive QDIO (OSA-Express) and iQDIO
(HiperSockets) adapters.
Alan Altmark
z/VM Development
IBM Endicott
--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO L
27;re
> faster and seem to be more reliable.
Don't confuse the DVD media with the CD media. z/VM 5.1 is available on
both. The CDs are for use only with a device that emulates a 3422 tape
drive (e.g. IBM 2074).
Alan Altmark
z/VM Development
IBM Endicott
, it is a terminal
control unit that allows you to telnet into it instead of having
hard-wired 3270s. The model 2 and later have a CD (not DVD) drive in them
that emulates a 3422 tape drive.
The DVD (not CD) installation media is used with the DVD drive in the HMC.
Alan Altmark
z/VM Develo
VM are on the same
> OSA. The gotcha there would be defining the right
> portname for an OSD if the OSA changes.
Of course, the OSAs no longer require portnames, so it is a non-issue.
Portnames are required only in z/OS configurations.
Alan Altmark
z/VM
ettings. And since IBM doesn't manufacture any of the switches these
days . If it works, great. If not, oh well.
Just add manual configuration to the problem determination checklist.
Alan Altmark
z/VM Development
IBM Endicott
---
d the educational
materials, and (c) are willing to do it for free, are few and far between.
:-)
SHARE is coming up at the end of February. That's a great place to get
z/VM education. In the DC/Maryland region, you can check out Hillgang,
the local z/VM and mainframe L
administration in a cogent fashion.
The z/VM hands-on labs at the zSeries Expo and at SHARE conferences have
been really popular. Those labs provide a *most excellent* structure for
presenting information.
Alan Altmark
z/VM Development
IBM Endicott
---
ice of CP level as an "innocuous
combination". Each release of z/VM since z/VM 4.2 has had focus on
improving the performance of Linux guests.
Just as with hardware, the choice of CP level should be a key element of a
oad, nothing happens unless I had already opened the 3270 integrated
> console window. Even with SYSG parameter for the Load, I still get the
IPL
> parameter screen (on 3270).
You need to run the SALIPL program to provide default pa
IG. E.g.
Operator_Consoles 020 SYSTEM_3270 SYSTEM_CONSOLE
Don't specify cons= on SALIPL; cons= will override OPERATOR_CONSOLES.
Alan Altmark
z/VM Development
IBM Endicott
--
For LINUX-390 subscribe / signoff / archive access ins
the delay when you deactivate the
LPAR (default is 0).
Linux will register for the LPAR deactivation signal just as it does on
VM.
Alan Altmark
z/VM Development
IBM Endicott
--
For LINUX-390 subscribe / signoff / archive ac
no
meaning.
It is not propagated to the real OSA and need not match any other portname
used on the same VSWITCH or Guest LAN.
Alan Altmark
z/VM Development
IBM Endicott
--
For LINUX-390 subscribe / signoff / archive access ins
nd if it's a
Hipersockets
> Guest LAN on z/VM, check the DEFINE LAN as that hardware setting is
mirrored
> there.
The value in HCD is the Maximum Frame Size (MFS). This is the OS= setting
for the chpid in the IOCDS. The MTU is al
uy
you anything, as (last I checked) the distros don't ship multiple
model-dependent copies of the kernel. And for some folks, recompiling the
kernel with different options could invalidate their service agreement.
Alan Altmark
z/VM Development
IBM Endicott
---
s; Uli Weigand has the best knowledge of the
processor-specific GCC changes.
Alan Altmark
z/VM Development
IBM Endicott
--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the mess
disk had performance implications. Not any more.
A group is a named collection of volids (volume names).
Every volume (and group, if you have them) must be in EXTENT CONTROL. You
can dynamically update EXTENT CONTROL by using DIRM SEND EXTENT CONTROL,
(receive and update
ure to be
dispatched could inhibit your ability to manage system performance.
For example, the TCPIP server is usually QUICKDSP as it needs to provide
interactive services to other virtual machines. If you need to telnet
into the system in order to fix a performance problem
Alan Altmark
z/V
has these numbers (except Max
interfaces) documented in the Machine Limits appendix.
Alan Altmark
z/VM Development
IBM Endicott
--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with
would encounter that same limitation if you use HCD on z/VM.
Feel free to let me know (offline, please) who in IBM gave you that
information and I can either (a) correct their understanding of how the
Universe functions or (b) lear
annels/addresses).
240 is indeed the OSD chpid device limit on z800/z900. (Increased to 480
on z890/z990.) That yields a maximum of 80 IP interfaces. If you've
defined an OSAD, then that reduces you to 79 IP interfaces.
> So, I don't know wher
quot; >/sys/devices/dcssblk/SWAPPING/save
> add this line to /etc/fstab: /dev/dcssblk0 swap swap defaults 0 0
> type swapon -a, and look at /proc/swaps to verify that everything worked
> fine
[insert]
And then shutdown this Linux and bring up your class G e
told you, your co-workers would have to kill me. :-)
Alan Altmark
z/VM Development
IBM Endicott
--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX
ication redundancy. Systems Automation for Linux
provides server failover using heartbeats and STOMITH (shoot the other
matchine in the head) technology.
Alan Altmark
z/VM Development
IBM Endicott
--
For LINUX-390 subscribe / si
ation will certainly do the
job,
> but at a much, much higher cost.
Do the open source solutions deal with STOMITH? SAfLi will perform FORCE
or will use LPAR management APIs to kill off the offending machine.
Alan Altmark
z/VM D
led because the dasd volume is not on a ESS (Shark). Now, why DIRMAINT
assumed FLASHCOPY would work is something someone else will have to
explain.
Alan Altmark
z/VM Development
IBM Endicott
--
For LINUX-390 subscribe / signoff
as just trying to illustrate that the
solution requires a different class of applications specifically targeted
to this need - that it isn't built into Linux. SAfLi is a commercial
offering in that space, if someone needs that level of function and
support. The open source options may fit the bill as
gt; different interfaces. Maybe you just need to write a VM plugin for it.
Now y'all cut that out. Use the z/VM System Management APIs; that's what
they're for. No need to give the HA monitors all that privilege.
Alan Altmark
z/VM Development
IBM Endicott
logoff
> ?
CPU 1 correctly loaded the 0xFFF signal acknowledgment PSW.
CPU 2 is properly in stopped state.
What happened to CPU 0? How many vCPUs are there?
Remove the vmpoff parameter from zipl.conf; SIGNAL SHUTDOWN and vmpoff are
mutually
Cables *do* break.
- Latest microcode.
- Have OSA/SF or remote HMC access available to you.
- Let your network techs do their job. Don't do it for them. (They can do
it better and faster than you can.)
Unfortunately you ran into the first three all at once.
Alan Altmark
z/VM
r. When the current OSA fails, the other OSA will be
used.
Alan Altmark
z/VM Development
IBM Endicott
--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-
nes defined for use by the VSWITCH. In 5.1 we make it harder for you to
break the rule. In 4.4 it is up to you.
Alan Altmark
z/VM Development
IBM Endicott
--
For LINUX-390 subscribe / signoff / archive access instructio
/ for the zOS TCP/IP stack? Is there a definition I'm missing?
Double check your IP addresses and subnet masks. The HiperSocket must be
treated like a separate LAN segment with its own IP subnet and unique IP
addresses. Packets may be making their way to z/OS, but the
code, and use IUCV or APPC instead of CP MSG
or SMSG, then you can use ISFC and Distributed IUCV (already part of your
system).
Alan Altmark
z/VM Development
IBM Endicott
--
For LINUX-390 subscribe / signoff / archive access instr
o subscriptions, etc.
A little bit of clarification: The IBM definition of an "engine" is an
IFL or standard CP that is configured in the CEC (capacity), not the
partition (usage).
Alan Altmark
z/VM Development
IBM Endicott
---
ultimate indignity, whatever cable we provide won't be the
right length.
That's the nice thing about the 1000baseT (z890/z990) card over the
Gigabit card: it uses standard RJ-45 connectors and CAT5 cables.
Alan Altmark
z/VM Development
IBM Endicott
-
y the switch/router will take care of network failures and the
VSWITCH will take care of switch or OSA failures.
Alan Altmark
z/VM Development
IBM Endicott
--
For LINUX-390 subscribe / signoff / archive access instructions,
send em
On Wednesday, 12/22/2004 at 11:42 ZE7, Andri W Sundara
<[EMAIL PROTECTED]> wrote:
> Hi,
> Is OSA-G giga bit supported by SNA ?
> thanks
The gigabit OSA does not support SNA (it is QDIO only). The 1000Base-T
supports both TCP/IP and SNA (it runs in QDIO and non-QDIO modes).
Al
primary on one switch is the backup on the
other.
Having the VSWITCH respond gracefully to high OSA utilization rates is on
our to-do list.
> OK I'll get of my soapbox, but remember there are a lot of people
working on
> Quagga solutions because it is not zSeries specific.....
It
plication
from interface outages. The good thing is that the users will not call if
you lose an IP interface. The bad thing is that the users will not call
you if you lose an IP interface. You have to depend on more traditional
network monitoring methods to notify you of an outage.
Alan
multiple
CPUs. Alternatives include installing z/VM 5.1 or using only one virtual
CPU.
Alan Altmark
Sr. Software Engineer
IBM z/VM Development
--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL
it is being used or not),
then vary it back online. That will cause the card to reset.
Without knowing more about your *intended* configuration, and information
about your *current* configuration, I can't help you. Sorry.
Alan Altmark
z/VM Development
IBM Endicott
-
er, not inside the mainframe. (Which is where dynamic routing
should be done, anyway, IMO.)
Alan Altmark
z/VM Development
IBM Endicott
--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PRO
quot;.
And hopefully you are using the dynamic I/O capabilities of z/VM or z/OS
so that you do not have to perform a POR.
Alan Altmark
z/VM Development
IBM Endicott
--
For LINUX-390 subscribe / signoff / archive access inst
data on
them). The server would use the Subsystem Device Driver (SDD) to access
the Shark and serve the data in whatever way you see fit.
Alan Altmark
z/VM Development
IBM Endicott
--
For LINUX-390 subscribe / signoff / archive
d. Now run samba on one of those servers and
you're done. (There's nothing *inside* the Shark that will run samba. It
has to be an external box of your choosing.)
Alan Altmark
z/VM Development
IBM Endicott
--
For LI
-disks") from Linux.
In addition to minidisks, NFS provides access to CMS files in the BFS and
SFS.
Alan Altmark
z/VM Development
IBM Endicott
--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAI
and change the physical switch port to a
trunk, everyone using that port must be VLAN-aware. You can't mix
VLAN-aware and VLAN-unaware traffic on the same port. (Well, you might be
able to in some cases, but you might also be unple
flashcopy via DDR (I was told), you need to access
> these mini disk but we can not. How does everyone else do it?
There is no "flashcopy via DDR". It is done by the CP FLASHCOPY command.
Alan Altmark
z/VM Development
IBM Endicott
--
t to make a file-system on the partition and
the device is ready for use (dasdfmt, fdasd, fdisk, don't work on FBA
devices).
You can also have a look into the Device Drivers book for more info on how
to get a FBA device working.
Alan Altmark
z/VM Development
IBM Endicott
--
y of refinement".
Are you not getting help when you call the problem in? (I hope people are
not treating Linux like a toaster and putting up with burnt toast. If the
folks who write the installation tools don't get service calls the
problems
bers
>
> and the /var/log/messages file notes:
>
> Jan 17 14:13:58 sles9 rpc.mountd: refused mount request from
> 192.168.101.110 for
> /root (/): illegal port 54519
Try specifying "LOCALPORT nn" on the mount, where "nn" is a port < 1024.
You will need to be in
ockets, you can either give each Linux guest a
connection to the real HiperSocket or give just one of the guests on the
VSWITCH a HiperSockets connection and route everyone else thorough that
guest.
Alan Altmark
z/VM Development
IBM Endicott
-
put into automation on z/OS.
>
> Nope. RSCS still supports direct physical CTC connections, and PROP
> still sends things via RSCS messages. The z/OS side gets them like any
> other remote message. No VM/VTAM required.
I think Rob is referring to the NetView family of monitor
ot;change is not reflected" you mean that
issuing "ifconfig hiperlec" shows the wrong (old) address?
I would be fascinated to see the TCPIP console log with TRACE SECURITY and
the output of NETSTAT DEV HOME. A call to the Support
I would like to take this opportunity to ask SHARE attendees to please
contribute a small portion of your time to help an organization that
represents YOUR interests.
I know, you're going "But SHARE is still a month away and I don't have
time to worry about it right now!" Well, other people have
never-ending battle to be able to find all the
pointers, and to detect the layout of the control blocks.
"z/VM gives you Infinite Power...the Wise Man treats it as he would water
in the desert."
Confucious, ca. 470 BC (ok ok...maybe it was Chuckie, I
don
often you check,
there's no additional [Linux] CPU required to make the check.
Alan Altmark
z/VM Development
IBM Endicott
--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the
ve not
done so already, so that the VIPA address is used on OSE1 and OSD3.
Alan Altmark
z/VM Development
IBM Endicott
--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with t
te or waiting for the DVD are your only options.
Alan Altmark
z/VM Development
IBM Endicott
--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390
rame implementation of anything you would recognize as "Unix", so the
implementation comparison is impossible.
To splitters of hairs: Yes, z/OS has a Unix-branded API and shell
available, but I you can't call z/OS "Unix".
Alan Altmark
z/VM Development
IBM Endicott
--
stion, while USS might pass the Unix "duck
test", the system as a whole would be alien to somone familiar with any of
the traditional Unix implementations.
This isn't a slight against USS, it's a recognition of the vast difference
between z/OS and
ys think of you guys in the context of your
contributions to Linux, and failed to recognize your other contributions.
Alan Altmark
z/VM Development
IBM Endicott
--
For LINUX-390 subscribe / signoff / archive access instructions
ve it can also
be used with FLEX-ES.
Its format is not suitable for use with the DVD drive in the HMC.
Alan Altmark
z/VM Development
IBM Endicott
--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to
ith https, authenticate at that
level, and then request REXECs.
Just pointing out that there are multiple ways to solve the problem. (I
happen to like SSH, but there are other solutions, too.)
Alan Altmark
z/VM Development
IBM Endicott
---
ets a bit muddied in the installation
doc. Thanks for pointing it out!
[breathe inbreathe out...inoutinhale energyexhale
stress]
Alan Altmark
z/VM Development
IBM Endicott
--
For LINUX-390 subscribe / signof
On Friday, 02/11/2005 at 05:22 CET, José Raúl Barón Rodriguez
<[EMAIL PROTECTED]> wrote:
> The link to that manual doesn't seem to work but thanks anyway
The URL was split. Try:
http://www10.software.ibm.com/developerworks/opensource/linux390/docu/lx24ju
n03dd04.pdf
Ala
ux and if I
NETSTAT
> DEV on zVM the packets received counter does not increment. Anyone seen
this
> before? Anyone have any ideas what I may need to fix this gigabit
ethernet
> interface?
Specify "PRIROUTER" in the TCP/IP DEVICE statement f
erating system you care to run on
it (with the provisos implied by the Architectural Level Set on zSeries
boxes).
Thank you for bringing this to our attention. We'll get the FAQ updated
ASAP PDQ.
Alan Altmark
z/VM Development
IBM Endicott
-
file system, BFS. It isn't required for SFS, minidisk, or RDR.
Make sure FTPSERVE is listed in the AUTOLOG statement in PROFILE TCPIP and
that there are no errors on TCPIP's console.
Alan Altmark
z/VM Development
IBM Endicott
ERVE CON WAS 0001 RECS 0007 CPY 001 T
NOHOLD
> NOKEEP
> VM TCP/IP Netstat Level 440
You need to view the console log (RDR file) that was sent. The FTP server
is trying to start, but is failing. The FTPSERVE console log will tell
you why. (Or you can just logon to FTPSERVE and watch it run
On Wednesday, 02/16/2005 at 10:11 CST, "Little, Chris"
<[EMAIL PROTECTED]> wrote:
> Ah. Chuckie has been busy lately . . . first APAR's, now FAQ's.
Now cut that out. Contrary to popular opinion, some planted fingerprints
(hello? Bourne Supremacy? duh! You think that stuff only happens in the
m
TF
for APAR VM63348 for z/VM V4.4 SCSI IPL support.
Alan Altmark
z/VM Development
IBM Endicott
--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INF
ntioned a lot in various Linux venues. It happens when
the app isn't enabled for > 2GB. Try a different FTP app.
It's likely that the ftp client is issuing a SIZE directive and the server
or client can't deal with it. (Loo
amic I/O
facilities, whether by command or by using HCD.
Or, you can simply connect VM TCP/IP to the VSWITCH as well.
Alan Altmark
z/VM Development
IBM Endicott
--
For LINUX-390 subscribe / signoff / archive access instructions,
. You could, in theory, point the Linux network install to
the VM FTP server and have it deliver the Linux files.
Alan Altmark
z/VM Development
IBM Endicott
--
For LINUX-390 subscribe / signoff / archive access instructions,
eployed application. If
> so, I'd say you have at least a working explanation of what's going on.
I would suggest reporting the "bloat" to the IBM Support Center. I think
I'd queue it to WebSphere and let
st?
The NICDEF directory statement enables you to specify the low-order 3
bytes of the MAC address. The VMLAN MACPREFIX statement in SYSTEM CONFIG
lets you specify the high-order 3 bytes. In z/VM 5.1 you can even specify
the range from which CP will assign dynamic vMACs.
Alan Altmark
z/VM
nable to run, say, Apache
and Tomcat on, it won't do for WebSphere. You have to get underneath the
kernel, so to speak, and find out what drives someone to install Linux on
the mainframe. That will usually drive the distro flavor.
Not in all cases, of course, but it's shown itself to
not wish to
participate in "early adopter" programs, and it isn't because they are
anti-Linux. The choice we were discussing was not "Linux vs. Not Linux",
but "FC vs. RHEL". Tim asked why FC would be considered unusual in the
mainframe, and
p later.
I would be surprised if NFS will work when FTP does not.
Alan Altmark
z/VM Development
IBM Endicott
--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LIN
On Friday, 03/11/2005 at 03:39 CET, Rob van der Heij <[EMAIL PROTECTED]>
wrote:
> PS Anyone know whether we need to spell YAST these days? It appears
> that Novell write SuSE Linux as SUSE LINUX. In a call yesterday one of
> my US peers suggested that they now can pronounce SUSE the American
> way,
RMS. (DEDICATE failures
don't appear in the console log since those messages occur before the
virtual console is ready.)
You may not have those devices at your disposal. It all depends on how
the OSA was defined in the IOCDS.
Alan Altmark
z/VM Development
IBM Endicott
-
r that they
will appear on the tape. If you REWIND 181 before you IPL, then the first
file the TDF points to must contain the bootable kernel.
There's nothing magic about the /tapes directory. It's just where the
files are located on the CDs we produce and the TDF point
program continues to read from the tape
starting at byte 25. Whether that's the same file or the next file in the
TDF is transparent to the channel program.
The different files that make up a CD can be in any directory on the CD.
It is the TDF that specifies the files.
m it.
You have to decide which kind of device you're going to be using: An
emulated tape or the HMC DVD drive. How you use them and how you build
them are wy different.
Alan Altmark
z/VM Development
IBM Endicott
e that the IPLable versions actually
load, I think.
Someone more conversant with Linux build will have to say where the binary
kernel image is stored these days.
Alan Altmark
z/VM Development
IBM Endicott
--
For LINUX-390
ve
is used for other purposes, too. Use the CD as a "recovery" system and
let your production system load from DASD.
Alan Altmark
z/VM Development
IBM Endicott
--
For LINUX-390 subscribe / signoff / archive access instru
HCP MSG'd a server that would
ATTACH the requested device (if authorized) to the requestor. Or follow
Neale's advice and move ATTACH to a different priv class and have PROP
monitor the operator's console to make sure the guest is attaching only
what you think it ought to.
Alan A
of patches to IUCV support since "April 2004"
stream was released.
Is there other IUCV activity going on (e.g. vmlogrdr)? How many virtual
CPUs? Of course, from a network perspective, it's time to dump IUCV and
move to z/VM Guest LANs or the Virtual Switch.
Alan Altmark
z/VM Deve
701 - 800 of 1955 matches
Mail list logo